Method and system for testing the integrity of filters

This is the U.S. National Stage of International Application No. PCT/EP2021/087500, filed Dec. 23, 2021, which was published in English under PCT Article 21(2), which in turn claims the benefit of European Application No. 20 217 390.2, filed Dec. 28, 2020. The prior applications are incorporated herein by reference in their entirety.

The following description relates to a method, a medium and a system for testing the integrity of filters in the pharmaceutical industry and/or the biotechnology industry.

Filter integrity is a fundamental element of sterility assurance during production of pharmaceutical (e.g. biopharmaceutical) and/or biotechnological products. Different types of integrity tests can be performed, including destructive and non-destructive tests. Non-destructive tests are particularly advantageous because they can be performed prior to the use of the filter. Examples of non-destructive tests include the diffusion test, the bubble point test and the water flow test (also called water intrusion test).

The underlying general concept for a non-destructive integrity test is as follows. An integrity tester pressurizes an external volume at the upstream side of the filter to a set test pressure and maintains said pressure for a duration defined by a stabilization time. After this, a quantity indicative of the integrity of the filter is determined during a check phase lasting for a duration defined by a check duration. If the quantity being measured stays below a predefined limit, the integrity test is evaluated to be a passed test. Thus, the stabilization time and the check duration affect the dependability and the efficiency of the test. Alternatively, direct flow measurements can also be performed, in which case a stabilization phase may not be required.

It is an object of the invention to improve the efficiency (in particular the time efficiency) of an integrity test while at the same time enhancing the dependability thereof.

The achievement of this object in accordance with the invention is set out in the independent claims. Further developments of the invention are the subject matter of the dependent claims.

According to one aspect, a method for testing integrity of a filter is provided. Said otherwise, a method for performing an integrity test of a filter is provided.

The filter may be any filter that is used in an industrial process in the biotechnological and/or biopharmaceutical field. For instance, the filter may be any one of the following: a depth filter, a pre-filter, a sterilizing grade filter, a mycoplasma retentive filter, a cross-flow (or tangential flow) filter, an ultrafiltration filter, a membrane adsorption filter, a virus retentive filter. The filter may be hydrophilic or hydrophobic. The filter may also be referred to as “filter assembly”.

Exemplarily, the filter may be a sterile membrane filter comprising a housing and a membrane inside the housing, the membrane having a given pore size, which may e.g. range from about 10 nm to about 5 μm. The membrane can be made for example of polyethersulfon, polyvinylidene fluoride, polytetrafluoroethylene, cellulose acetate, regenerated cellulose and nylon. The housing can be made for example from polypropylene, polyamide or polytetrafluoroethylene.

The integrity test may be used to check for cracks and other defects in the filter that would compromise its functionality. The integrity of the filter may be tested applying the principles of established non-destructive techniques, such as the diffusion test (including a multipoint diffusion test) or the water flow test. These tests rely on the correlation between a physical quantity that can be easily determined from measurements and the actual retention capability of the filter. In the diffusion test, the physical quantity is the diffusion of a gas through a wetted filter, while in the water flow test the physical quantity is the flow of water through a hydrophobic filter.

The method comprises pressurizing an upstream side of the filter to a test pressure. In other words, the method comprises increasing the pressure at the upstream side of the filter until the pressure has reached a predetermined or predeterminable value, i.e. the test pressure.

A filter has an upstream side, which is the feed side of the filter, i.e. the surface of the filter through which the feed passes in order to be filtered, and a downstream side, which is the filtrate side of the filter, i.e. the surface of the filter from which the filtrate comes out after part of the feed is retained by the filter. For example, if the filter comprises a cylindrical membrane, the upstream side may be the outer surface of the cylinder and the downstream side may be the inner surface of the cylinder.

The pressure at the upstream side of the filter corresponds to the pressure that can be measured in an enclosed volume that is delimited, among others, by the upstream side of the filter, which is referred to as upstream volume. The specifics of this volume depend on the type of filter and on the configuration of the integrity tester. Exemplarily, for a filter comprising a housing, the upstream volume may be given by the net volume of the housing, the volume of connecting tubes and of the elements connected to the housing (such as the integrity tester).

The upstream volume is pressurized by introducing gas into the volume, e.g. by means of a gas inlet line connecting the integrity tester and the filter. The gas may be e.g. compressed air, carbon dioxide, nitrogen. The test pressure is the pressure that has to be reached for performing the integrity test. Indeed, there needs to be a pressure difference between the upstream side of the filter and the downstream side of the filter in order for a fluid to go through the filter to “test” it. A test fluid may be a liquid or a gas (pure gas or gas mixture). The test pressure pmay be in the range between about 25 mbarg and about 10 barg. The test pressure is a gauge pressure measured in relation to ambient atmospheric pressure.

The pressurization is one of the phases of the integrity test. The following phase is the check phase, optionally preceded by a stabilization phase. The check phase is the actual integrity checking part of the test, in which it is determined whether the filter is integral or not.

Indeed, the method further comprises performing a check step. The check step comprises determining a rate of the test fluid from the upstream side to a downstream side of the filter. The flow rate of the test fluid is the amount (or volume) of fluid that passes through the filter in a unit of time. The fluid may be a gas and the flow rate may be a diffusional flow rate. This is e.g. the case when the integrity test is a diffusion test, which can be performed on hydrophilic filters. Alternatively, the test fluid may be a liquid, e.g. water, and the flow rate may be a bulk (or volumetric) flow rate. This is e.g. the case when the integrity test is a water flow test (or water intrusion test), which can be performed on hydrophobic filters.

Depending on the specific integrity test, the filter may have to be prepared before the test begins. For instance, a hydrophilic filter may be wetted with wetting liquid, such as water or a mixture of water and alcohol, while a hydrophobic filter may be wetted with isopropyl alcohol.

Determining the flow rate may comprise, depending on the methodology employed to do so, measuring one or more physical quantities and/or performing calculations. For example, determining the flow rate may comprise measuring a pressure drop.

During the check phase, the pressure at the upstream side of the filter drops as the fluid diffuses through the filter. If p(t) is the measured instantaneous pressure at a given time point twhich is separated by a time interval δfrom the start of the check phase, which may substantially coincide with the end of the pressurization phase, or optionally the end of a stabilization phase as discussed below, a pressure drop can be defined as Δp=p−p, where pis the pressure at the start of the check phase.

In the context of a diffusion test, the flow rate F(t)=D(t) at the time point tmay exemplarily be determined as follows:

The upstream volume may be given a priori (e.g. retrieved from a storage medium or input by a user) or, in a particular example, the method may further comprise measuring the upstream volume, i.e. the volume at the upstream side of the filter. The determination of the upstream volume can be carried out according to conventional techniques, e.g. using Boyle's law, or may be input by a user.

In the context of a water flow test, the flow rate F(t)=B(t) at the time point tmay exemplarily be determined as follows:

Other ways of determining the flow rate may be used.

The check step further comprises comparing the determined flow rate with a flow range including a flow threshold. The flow range is a range of values defined by a lower limit Fand an upper limit For, said otherwise, is the set of values containing all values between the lower limit and the upper limit. The flow range may be an open interval or a closed interval. The flow threshold Fis a threshold value and belongs to the flow range, i.e. F<F<F, or F=F−Cand F=F+Cwith C, Cconstants. In some examples, the flow threshold may be the midpoint of the flow range, i.e. C=C.

The flow threshold is an upper limit for the flow rate: a filter exhibiting flow rates higher than this upper limit may be declared as non-integral. The value of the flow threshold may depend on a plurality of factors including, but not limited to, the type of integrity test, the area of the filter, the thickness of the filter, the test pressure. For example, for a diffusion test the flow threshold may range from about 1-3 ml/mi to about 200-300 ml/min, depending on the characteristics of the filter. Exemplarily, the value of the flow threshold may be assessed by performing a bacterial challenge test and it may be provided in the technical specifications of a filter.

The lower limit and the upper limit are determined accordingly to create an interval containing the flow threshold and their specific values may be set based on at least some of the factors listed above. Exemplarily Fmay be at least 85% of F, further exemplarily at least 90% of F, yet further exemplarily 92% of F. Exemplarily Fmay be at most 115% of F, further exemplarily at most 110% of F, yet further exemplarily 108% of F.

The flow range may exemplarily be stored in a storage medium, so that comparing the determined flow rate with the flow range may comprise retrieving the flow range from the storage medium.

The check step further comprises setting stop criteria based on the comparison and determining whether the stop criteria are satisfied. If the stop criteria are not satisfied, the method comprises repeating the check step until the stop criteria are satisfied.

The stop criteria are conditions that allow to determine whether the determined flow rate is sufficiently reliable, i.e. whether it provides a sufficiently accurate estimate of the actual flow rate. Indeed, as for all physical quantities, the true value (the actual flow rate) cannot be known but only estimated. In particular, since the flow rate is derived from measured quantities, the determined flow rate is affected by errors inherent to the measuring process, such as uncertainties and/or artefacts.

Accordingly, the determined flow rate at a given time point may not accurately reflect the actual flow rate. However, the actual flow rate is related to the retention capability of the filter, which is the object of the integrity test. Therefore, in order to correctly assess the integrity of the filter, it should be ensured that the determined flow rate is reliable enough.

Thus, the check step (particularly the determination of the flow rate) is repeated over time until it is established that the determined flow rate is sufficiently accurate, namely that the stop criteria are satisfied. Hence, the check phase may comprise one or more iterations of the check step and, further, it comprises determining whether the filter is integral.

Each repetition of the check step may occur after a predetermined or predeterminable time interval from the previous execution of the check step, so that the first check step is performed at time t, the second check step at time t=t+Δt, the third check step at time t=t+Δt+Δtand so on. Exemplarily, the time intervals between the iterations may have varying duration, e.g. have decreasing duration as time passes (Δt>Δt), or they may be constant (Δt=Δt). The durations of the time intervals may be fixed constants or may depend on one or more conditions being satisfied.

Once it is determined that the stop criteria are satisfied, the reiteration of determining the flow rate can be stopped. The period of time between the start of the check phase and the last iteration of the check step, i.e. the duration of the check phase, may be denoted as check duration. The end time point of the check phase may be denoted as check end time.

Hence, the repetition of the check step leads to a time series, i.e. a series of data points corresponding to different times, wherein each data point is a determined flow rate F(t), with i∈[1,n] and tbeing the check end time. Each determined flow rate corresponds to a different time point and a different iteration of the check step.

Conventionally, the check duration is a predetermined, fixed quantity that is set to be long enough to ensure a reliable determined flow rate. According to the invention, the check duration is not a fixed quantity and depends instead on when the stop criteria are met. In other words, the check duration is a variable contingent on the actual reliability of the determined flow rate as prescribed by the stop criteria.

This leads to more time-efficient and more accurate integrity testing as compared to a fixed check duration. Indeed, if the determined flow rate is reliable at a time before the fixed check end time, the variable check duration will be shorter than the fixed check duration. If the determined flow rate becomes reliable at a time after the fixed check end time, the result of the test will be more accurate with the variable check duration.

Optionally, the stop criteria may be supplemented by a maximum check duration, which is a value that may be pre-programmed or may be set by a user. If the duration of the check phase reaches the maximum check duration, the repetition of the check step may be stopped even if the stop criteria have not yet been met. In this case, the last determined flow rate may be compared with the flow threshold, as discussed below.

Furthermore, based on the comparison between the determined flow rate and the flow range, the stop criteria are differently set. Specifically, if the determined flow rate is within the flow range, the stop criteria are set to first stop criteria; and if the determined flow rate is outside the flow range, the stop criteria are set to second stop criteria.

The second stop criteria are different from the first stop criteria. In particular, the second stop criteria are less stringent than the first stop criteria. This means that a determined flow rate that satisfies the second stop criteria does not necessarily satisfy the first stop criteria, while a determined flow rate that satisfies the first stop criteria always satisfies the second stop criteria.

Accordingly, the determined flow rate at a given time point may satisfy only the second stop criteria but not the first stop criteria. Thus, the check duration is different depending on how the stop criteria have been set, specifically it may be shorter if the stop criteria are the second stop criteria, thereby making the test faster but not less accurate.

The choice of the stop criteria and, consequently, of the check duration is based on the comparison of the determined flow rate with the flow range. In other words, the value of the determined flow rate is checked against the flow range to verify whether the determined flow rate F(t) lies within the flow range or outside the flow range, i.e. whether F(t)∈[F,F] or not.

Since the flow range is an interval around the flow threshold, if the determined flow rate lies within the flow range, it is closer to the flow threshold, while if it lies outside the flow range, it is farther away from the flow threshold. As will be discussed below, in order to assess the integrity of the filter, it is checked whether the determined flow rate exceeds the flow threshold or not.

Since the relation of the determined flow rate to the flow threshold is decisive, when the determined flow rate is near the flow threshold the accuracy of the determined flow rate is of particular importance. Indeed, even a relatively smaller inaccuracy in the determined flow rate may lead, in this case, to a wrong assessment about the filter integrity. Conversely, if the determined flow rate is sufficiently distant from the flow threshold (i.e. outside the flow range), even a relatively larger inaccuracy would not change the result of the assessment.

Accordingly, the stop criteria are stricter when the determined flow rate is within the flow range and less stringent when the determined flow rate is outside the flow range. The check duration is, thus, adaptively determined: it is longer when a more careful scrutiny is required and it is shorter when the determined flow rate is already sufficiently robust for the comparison with the flow threshold. Therefore the method enhances the time efficiency of an integrity test while maintaining its accuracy, in other words without comprising its quality.

In a particular example, the flow range may be a first flow range and the stop criteria may be set to the second stop criteria if, further to the determined flow range being outside the first flow range, the determined flow range is within a second flow range, the first flow range being included in the second flow range; and the method may further comprise setting the stop criteria to third stop criteria if the determined flow rate is outside the second flow range.

More generally, there may be more than two different stop criteria, wherein the stop criteria are defined to be progressively less stringent as the determined flow rate gets farther and farther away from the flow threshold. In this case, a plurality of flow ranges may be defined, wherein the number of different stop criteria corresponds to the number of flow ranges plus one.

Each flow range may be around (e.g. centered on) the flow threshold: a first flow range [F,F] may be the smallest range around F, a second flow range [F, F] may include the first flow range [F,F]⊃[F,F], a third flow range may include the second flow range [F,F]⊃[F,F] and so on. Accordingly, F∈[F,F]⊂[F,F]⊂[F, F]⊂ . . . ⊂[F,F] for n flow ranges.

Therefore, the stop criteria may be set as follows: